1 paper · 1 filter
Julius Porbeck, Christian Medeiros Adriano, Holger Giese
Large language model (LLM)-based debugging systems can generate failure explanations, but these explanations may be incomplete or incorrect. Misleading explanations are harmful for…